vga_is_firmware_default() checks firmware resources to find the owner
framebuffer resources to find the firmware PCI device. This is an
open coded implementation of screen_info_pci_dev(). Switch to using
screen_info_pci_dev() instead.
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Suggested-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
---
v10:
* Rebase on 6.17-rc1
---
drivers/pci/vgaarb.c | 31 +++++--------------------------
1 file changed, 5 insertions(+), 26 deletions(-)
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 78748e8d2dbae..b58f94ee48916 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -556,34 +556,13 @@ EXPORT_SYMBOL(vga_put);
static bool vga_is_firmware_default(struct pci_dev *pdev)
{
-#if defined(CONFIG_X86)
- u64 base = screen_info.lfb_base;
- u64 size = screen_info.lfb_size;
- struct resource *r;
- u64 limit;
+#ifdef CONFIG_SCREEN_INFO
+ struct screen_info *si = &screen_info;
- /* Select the device owning the boot framebuffer if there is one */
-
- if (screen_info.capabilities & VIDEO_CAPABILITY_64BIT_BASE)
- base |= (u64)screen_info.ext_lfb_base << 32;
-
- limit = base + size;
-
- /* Does firmware framebuffer belong to us? */
- pci_dev_for_each_resource(pdev, r) {
- if (resource_type(r) != IORESOURCE_MEM)
- continue;
-
- if (!r->start || !r->end)
- continue;
-
- if (base < r->start || limit >= r->end)
- continue;
-
- return true;
- }
-#endif
+ return pdev == screen_info_pci_dev(si);
+#else
return false;
+#endif
}
static bool vga_arb_integrated_gpu(struct device *dev)
--
2.43.0Hi, On Mon, Aug 11, 2025 at 11:26:04AM -0500, Mario Limonciello (AMD) wrote: > vga_is_firmware_default() checks firmware resources to find the owner > framebuffer resources to find the firmware PCI device. This is an > open coded implementation of screen_info_pci_dev(). Switch to using > screen_info_pci_dev() instead. > > Acked-by: Bjorn Helgaas <bhelgaas@google.com> > Suggested-by: Thomas Zimmermann <tzimmermann@suse.de> > Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org> I'm getting a black screen on boot on mainline, and it bisected to this commit. Reverting this commit fixed it. Please revert. - Eric
[+cc regressions]
On Sun, Oct 12, 2025 at 11:23:02AM -0700, Eric Biggers wrote:
> On Mon, Aug 11, 2025 at 11:26:04AM -0500, Mario Limonciello (AMD) wrote:
> > vga_is_firmware_default() checks firmware resources to find the owner
> > framebuffer resources to find the firmware PCI device. This is an
> > open coded implementation of screen_info_pci_dev(). Switch to using
> > screen_info_pci_dev() instead.
> >
> > Acked-by: Bjorn Helgaas <bhelgaas@google.com>
> > Suggested-by: Thomas Zimmermann <tzimmermann@suse.de>
> > Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
>
> I'm getting a black screen on boot on mainline, and it bisected to this
> commit. Reverting this commit fixed it.
#regzbot introduced: 337bf13aa9dd ("PCI/VGA: Replace vga_is_firmware_default() with a screen info check")
#regzbot link: https://lore.kernel.org/r/20251013154441.1000875-1-superm1@kernel.org
On Mon, Oct 13, 2025 at 03:55:31PM -0500, Bjorn Helgaas wrote:
> On Sun, Oct 12, 2025 at 11:23:02AM -0700, Eric Biggers wrote:
> > On Mon, Aug 11, 2025 at 11:26:04AM -0500, Mario Limonciello (AMD) wrote:
> > > vga_is_firmware_default() checks firmware resources to find the owner
> > > framebuffer resources to find the firmware PCI device. This is an
> > > open coded implementation of screen_info_pci_dev(). Switch to using
> > > screen_info_pci_dev() instead.
> > >
> > > Acked-by: Bjorn Helgaas <bhelgaas@google.com>
> > > Suggested-by: Thomas Zimmermann <tzimmermann@suse.de>
> > > Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
> >
> > I'm getting a black screen on boot on mainline, and it bisected to this
> > commit. Reverting this commit fixed it.
>
> #regzbot introduced: 337bf13aa9dd ("PCI/VGA: Replace vga_is_firmware_default() with a screen info check")
> #regzbot link: https://lore.kernel.org/r/20251013154441.1000875-1-superm1@kernel.org
#regzbot fix: a78835b86a44 ("PCI/VGA: Select SCREEN_INFO on X86")
On 10/12/25 1:23 PM, Eric Biggers wrote: > Hi, > > On Mon, Aug 11, 2025 at 11:26:04AM -0500, Mario Limonciello (AMD) wrote: >> vga_is_firmware_default() checks firmware resources to find the owner >> framebuffer resources to find the firmware PCI device. This is an >> open coded implementation of screen_info_pci_dev(). Switch to using >> screen_info_pci_dev() instead. >> >> Acked-by: Bjorn Helgaas <bhelgaas@google.com> >> Suggested-by: Thomas Zimmermann <tzimmermann@suse.de> >> Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org> > > I'm getting a black screen on boot on mainline, and it bisected to this > commit. Reverting this commit fixed it. > > Please revert. > > - Eric Can you please share more information about your issue before we jump straight into a revert? What kind of hardware do you have? Perhaps a kernel log from mainline and another from mainline with the revert could help identify what's going on? A revert might be the right solution, but I would rather fix the issue if it's plausible to do so.
On Sun, Oct 12, 2025 at 01:37:33PM -0500, Mario Limonciello wrote:
>
>
> On 10/12/25 1:23 PM, Eric Biggers wrote:
> > Hi,
> >
> > On Mon, Aug 11, 2025 at 11:26:04AM -0500, Mario Limonciello (AMD) wrote:
> > > vga_is_firmware_default() checks firmware resources to find the owner
> > > framebuffer resources to find the firmware PCI device. This is an
> > > open coded implementation of screen_info_pci_dev(). Switch to using
> > > screen_info_pci_dev() instead.
> > >
> > > Acked-by: Bjorn Helgaas <bhelgaas@google.com>
> > > Suggested-by: Thomas Zimmermann <tzimmermann@suse.de>
> > > Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
> >
> > I'm getting a black screen on boot on mainline, and it bisected to this
> > commit. Reverting this commit fixed it.
> >
> > Please revert.
> >
> > - Eric
>
> Can you please share more information about your issue before we jump
> straight into a revert? What kind of hardware do you have? Perhaps a
> kernel log from mainline and another from mainline with the revert could
> help identify what's going on?
>
> A revert might be the right solution, but I would rather fix the issue if
> it's plausible to do so.
Relevant hardware is:
AMD Ryzen 9 9950X 16-Core Processor
Radeon RX 9070
The following message appears in the good log but not the bad log:
fbcon: amdgpudrmfb (fb0) is primary device
I don't have CONFIG_SCREEN_INFO enabled, so the commit changed
vga_is_firmware_default() to always return false.
If DRM_AMDGPU depends on SCREEN_INFO now, it needs to select it.
- Eric
On 10/12/25 1:47 PM, Eric Biggers wrote:
> On Sun, Oct 12, 2025 at 01:37:33PM -0500, Mario Limonciello wrote:
>>
>>
>> On 10/12/25 1:23 PM, Eric Biggers wrote:
>>> Hi,
>>>
>>> On Mon, Aug 11, 2025 at 11:26:04AM -0500, Mario Limonciello (AMD) wrote:
>>>> vga_is_firmware_default() checks firmware resources to find the owner
>>>> framebuffer resources to find the firmware PCI device. This is an
>>>> open coded implementation of screen_info_pci_dev(). Switch to using
>>>> screen_info_pci_dev() instead.
>>>>
>>>> Acked-by: Bjorn Helgaas <bhelgaas@google.com>
>>>> Suggested-by: Thomas Zimmermann <tzimmermann@suse.de>
>>>> Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
>>>
>>> I'm getting a black screen on boot on mainline, and it bisected to this
>>> commit. Reverting this commit fixed it.
>>>
>>> Please revert.
>>>
>>> - Eric
>>
>> Can you please share more information about your issue before we jump
>> straight into a revert? What kind of hardware do you have? Perhaps a
>> kernel log from mainline and another from mainline with the revert could
>> help identify what's going on?
>>
>> A revert might be the right solution, but I would rather fix the issue if
>> it's plausible to do so.
>
> Relevant hardware is:
> AMD Ryzen 9 9950X 16-Core Processor
> Radeon RX 9070
>
> The following message appears in the good log but not the bad log:
>
> fbcon: amdgpudrmfb (fb0) is primary device
>
> I don't have CONFIG_SCREEN_INFO enabled, so the commit changed
> vga_is_firmware_default() to always return false.
Thanks, that definitely explains it.
>
> If DRM_AMDGPU depends on SCREEN_INFO now, it needs to select it.
>
> - Eric
Well the question now is which driver should actually select it.
Although it manifested for you in amdgpu, I don't think this is going to
be an amdgpu unique issue.
Maybe this:
diff --git a/drivers/video/fbdev/core/Kconfig
b/drivers/video/fbdev/core/Kconfig
index 006638eefa41..ce2544924b0e 100644
--- a/drivers/video/fbdev/core/Kconfig
+++ b/drivers/video/fbdev/core/Kconfig
@@ -5,6 +5,7 @@
config FB_CORE
select VIDEO
+ select SCREEN_INFO
tristate
config FB_NOTIFY
Hi Mario, thanks for handling this report quickly. Am 12.10.25 um 21:06 schrieb Mario Limonciello: [...] >> I don't have CONFIG_SCREEN_INFO enabled, so the commit changed >> vga_is_firmware_default() to always return false. > > Thanks, that definitely explains it. > >> >> If DRM_AMDGPU depends on SCREEN_INFO now, it needs to select it. >> >> - Eric > > Well the question now is which driver should actually select it. > > Although it manifested for you in amdgpu, I don't think this is going > to be an amdgpu unique issue. > > Maybe this: > > diff --git a/drivers/video/fbdev/core/Kconfig > b/drivers/video/fbdev/core/Kconfig > index 006638eefa41..ce2544924b0e 100644 > --- a/drivers/video/fbdev/core/Kconfig > +++ b/drivers/video/fbdev/core/Kconfig > @@ -5,6 +5,7 @@ > > config FB_CORE > select VIDEO > + select SCREEN_INFO Let's select it from CONFIG_VGA_ARB. This will restore the old behavior for x86. The else branch for CONFIG_SCREEN_INFO can then go away. Best regards Thomas > tristate > > config FB_NOTIFY -- -- Thomas Zimmermann Graphics Driver Developer SUSE Software Solutions Germany GmbH Frankenstrasse 146, 90461 Nuernberg, Germany GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman HRB 36809 (AG Nuernberg)
Am 11.08.25 um 18:26 schrieb Mario Limonciello (AMD):
> vga_is_firmware_default() checks firmware resources to find the owner
> framebuffer resources to find the firmware PCI device. This is an
> open coded implementation of screen_info_pci_dev(). Switch to using
> screen_info_pci_dev() instead.
>
> Acked-by: Bjorn Helgaas <bhelgaas@google.com>
> Suggested-by: Thomas Zimmermann <tzimmermann@suse.de>
> Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
> ---
> v10:
> * Rebase on 6.17-rc1
> ---
> drivers/pci/vgaarb.c | 31 +++++--------------------------
> 1 file changed, 5 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
> index 78748e8d2dbae..b58f94ee48916 100644
> --- a/drivers/pci/vgaarb.c
> +++ b/drivers/pci/vgaarb.c
> @@ -556,34 +556,13 @@ EXPORT_SYMBOL(vga_put);
>
> static bool vga_is_firmware_default(struct pci_dev *pdev)
> {
> -#if defined(CONFIG_X86)
> - u64 base = screen_info.lfb_base;
> - u64 size = screen_info.lfb_size;
> - struct resource *r;
> - u64 limit;
> +#ifdef CONFIG_SCREEN_INFO
> + struct screen_info *si = &screen_info;
>
> - /* Select the device owning the boot framebuffer if there is one */
> -
> - if (screen_info.capabilities & VIDEO_CAPABILITY_64BIT_BASE)
> - base |= (u64)screen_info.ext_lfb_base << 32;
> -
> - limit = base + size;
> -
> - /* Does firmware framebuffer belong to us? */
> - pci_dev_for_each_resource(pdev, r) {
> - if (resource_type(r) != IORESOURCE_MEM)
> - continue;
> -
> - if (!r->start || !r->end)
> - continue;
> -
> - if (base < r->start || limit >= r->end)
> - continue;
> -
> - return true;
> - }
> -#endif
> + return pdev == screen_info_pci_dev(si);
> +#else
> return false;
> +#endif
> }
>
> static bool vga_arb_integrated_gpu(struct device *dev)
--
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nuernberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman
HRB 36809 (AG Nuernberg)
© 2016 - 2026 Red Hat, Inc.